A LC Model With Change-Points Regime

Abstract

This paper extends the widely used Lee Carter (LC) model (Lee & Carter, 1992) for mortality projection. We suggest a Bayesian change-points model for the time parameters in the Bayesian extension of the LC model suggested in Czado et al. (2005). In particular, we modify the simple linear trend to a piecewise linear trend. This model accounts for changes in trend over time and it is inspired by the Bayesian random level{shift model of McCulloch & Tsay (1993). In a validation-based examination, the proposed change-points model produces smaller prediction errors compared to the autoregressive model for the time parameters in Czado et al. (2005). Notably, this is true for all populations considered.